Accurate path loss prediction is critical for the efficient planning and optimization of 5G and emerging 6G wireless networks, particularly in high-frequency millimeter wave (mmWave) bands. Traditional empirical models are limited in their ability to capture the complex and nonlinear characteristics of modern urban propagation environments. This paper proposes a hybrid machine learning framework that combines Random Forest, XGBoost, and deep neural networks to enhance prediction accuracy. The model utilizes a comprehensive set of input features, including distance, frequency, antenna heights, building density, and line-of-sight conditions, derived from a deterministic ray-tracing dataset. A weighted ensemble strategy is introduced to integrate the strengths of tree-based and deep learning models, enabling effective modeling of both discontinuous shadowing effects and smooth signal variations. Experimental results demonstrate that the proposed approach significantly outperforms classical models and individual machine learning methods, achieving an RMSE of 2.5 dB and an R² of 0.96. The results confirm the effectiveness of hybrid AI-based models for accurate path loss prediction and highlight their potential for next-generation wireless network design and optimization.